Section Article

An Innovative Method for Retrieving Content-Based Images By using colour characteristics
Author(s): Dr. Aalok Mishra

Abstract
Local colour histograms are well recognised as a simple and often used technique for doing color-based image retrieval. However there is a substantial quantity of space located above it. The implementation of a cell grid overlay on the images together with the use of histograms for each colour collectively result in a decrease in the necessary storage capacity. Therefore the approach in issue becomes responsive to the rotation and translation of pictures. The aim of this work is to provide a novel approach for representing the similarity between pictures which we refer to as HG. This approach uses a weighted undirected network for each colour where each vertex of the graph corresponds to a bin in the histogram distribution. Consequently the approach may use colour location and requires little additional space. However it is not sensitive to rotation or translation. We choose to carry out an experiment with a collection of 9600 distinct landscape images in a library. Our experimental findings demonstrate that our approach is superior in terms of efficiency compared to both the cell/color histogram and the local colour histogram.